Files
baoyu-skills/skills/baoyu-format-markdown/scripts/cjk-emphasis.ts
T

315 lines
8.4 KiB
TypeScript
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
const CJK_PUNCT_COMMON = "。.,、?!:;";
const CJK_OPENING_PUNCT = "(〔〖〘〚「『〈《【\u201C\u2018";
const CJK_CLOSING_PUNCT = ")〕〗〙〛」』〉》】\u201D\u2019" + CJK_PUNCT_COMMON;
const CJK_SCRIPTS =
"\\p{Script=Han}\\p{Script=Hiragana}\\p{Script=Katakana}\\p{Script=Hangul}";
export const CJK_CLOSING_PUNCT_RE = new RegExp(`[${CJK_CLOSING_PUNCT}]`);
export const CJK_OPENING_PUNCT_RE = new RegExp(`^[${CJK_OPENING_PUNCT}]`);
export const CJK_CHAR_RE = new RegExp(`[${CJK_SCRIPTS}]`, "u");
const PUNCT_OR_SYMBOL_RE = /[\p{P}\p{S}]/u;
const WORD_CHAR_RE = /[\p{L}\p{N}]/u;
const CJK_PUNCT_PAIRS: Record<string, string> = {
"“": "”",
"": "",
"": "",
"": "",
"〖": "〗",
"〘": "〙",
"〚": "〛",
"「": "」",
"『": "』",
"〈": "〉",
"《": "》",
"【": "】",
};
function findInlineCodeRanges(text: string): Array<[number, number]> {
const ranges: Array<[number, number]> = [];
let i = 0;
while (i < text.length) {
if (text[i] !== "`") {
i += 1;
continue;
}
let run = 1;
while (i + run < text.length && text[i + run] === "`") {
run += 1;
}
const start = i;
let j = i + run;
let found = false;
while (j < text.length) {
if (text[j] !== "`") {
j += 1;
continue;
}
let closeRun = 1;
while (j + closeRun < text.length && text[j + closeRun] === "`") {
closeRun += 1;
}
if (closeRun === run) {
ranges.push([start, j + closeRun - 1]);
i = j + closeRun;
found = true;
break;
}
j += closeRun;
}
if (!found) {
i = start + run;
}
}
return ranges;
}
function isEscaped(text: string, pos: number): boolean {
let count = 0;
for (let i = pos - 1; i >= 0 && text[i] === "\\"; i -= 1) {
count += 1;
}
return count % 2 === 1;
}
function isWhitespaceChar(ch: string | undefined): boolean {
return !ch || /\s/u.test(ch);
}
function isPunctuationOrSymbol(ch: string | undefined): boolean {
return !!ch && PUNCT_OR_SYMBOL_RE.test(ch);
}
function isMatchingCjkPunct(open: string, close: string): boolean {
return CJK_PUNCT_PAIRS[open] === close;
}
function mapNonCodeSegments(
block: string,
mapper: (segment: string) => string
): string {
const codeRanges = findInlineCodeRanges(block);
if (codeRanges.length === 0) {
return mapper(block);
}
let result = "";
let lastIndex = 0;
for (const [start, end] of codeRanges) {
if (start > lastIndex) {
result += mapper(block.slice(lastIndex, start));
}
result += block.slice(start, end + 1);
lastIndex = end + 1;
}
if (lastIndex < block.length) {
result += mapper(block.slice(lastIndex));
}
return result;
}
function moveCjkPunctuationOutsideEmphasis(block: string): string {
return mapNonCodeSegments(block, (segment) => {
const delimiterPositions: number[] = [];
let cursor = 0;
while (cursor < segment.length - 1) {
if (
segment[cursor] === "*" &&
segment[cursor + 1] === "*" &&
segment[cursor - 1] !== "*" &&
segment[cursor + 2] !== "*" &&
!isEscaped(segment, cursor)
) {
delimiterPositions.push(cursor);
cursor += 2;
continue;
}
cursor += 1;
}
if (delimiterPositions.length < 2) return segment;
const stack: number[] = [];
const pairs: Array<{ open: number; close: number }> = [];
for (const pos of delimiterPositions) {
if (stack.length === 0) {
stack.push(pos);
} else {
const open = stack.pop() as number;
pairs.push({ open, close: pos });
}
}
const skip = new Set<number>();
const insertBefore = new Map<number, string>();
for (const pair of pairs) {
const openPunctPos = pair.open + 2;
const closePunctPos = pair.close - 1;
if (openPunctPos >= closePunctPos) continue;
const openPunct = segment[openPunctPos];
const closePunct = segment[closePunctPos];
if (!openPunct || !closePunct) continue;
if (!CJK_OPENING_PUNCT_RE.test(openPunct)) continue;
if (!CJK_CLOSING_PUNCT_RE.test(closePunct)) continue;
if (!isMatchingCjkPunct(openPunct, closePunct)) continue;
if (openPunctPos + 1 >= closePunctPos) continue;
const inner = segment.slice(openPunctPos + 1, closePunctPos);
if (inner.length === 0) continue;
skip.add(openPunctPos);
skip.add(closePunctPos);
insertBefore.set(pair.open, (insertBefore.get(pair.open) ?? "") + openPunct);
const afterClose = pair.close + 2;
insertBefore.set(
afterClose,
(insertBefore.get(afterClose) ?? "") + closePunct
);
}
if (skip.size === 0) return segment;
let result = "";
for (let idx = 0; idx < segment.length; idx += 1) {
const insert = insertBefore.get(idx);
if (insert) {
result += insert;
}
if (skip.has(idx)) {
continue;
}
result += segment[idx];
}
const tailInsert = insertBefore.get(segment.length);
if (tailInsert) {
result += tailInsert;
}
return result;
});
}
function fixCjkEmphasisSpacingInBlock(block: string): string {
const normalized = moveCjkPunctuationOutsideEmphasis(block);
const codeRanges = findInlineCodeRanges(normalized);
let rangeIndex = 0;
const delimiters: Array<{
pos: number;
canOpen: boolean;
canClose: boolean;
}> = [];
let cursor = 0;
while (cursor < normalized.length - 1) {
if (rangeIndex < codeRanges.length && cursor >= codeRanges[rangeIndex][0]) {
if (cursor <= codeRanges[rangeIndex][1]) {
cursor = codeRanges[rangeIndex][1] + 1;
continue;
}
rangeIndex += 1;
continue;
}
if (
normalized[cursor] === "*" &&
normalized[cursor + 1] === "*" &&
normalized[cursor - 1] !== "*" &&
normalized[cursor + 2] !== "*" &&
!isEscaped(normalized, cursor)
) {
const before = normalized[cursor - 1];
const after = normalized[cursor + 2];
const beforeIsSpace = isWhitespaceChar(before);
const afterIsSpace = isWhitespaceChar(after);
const beforeIsPunct = isPunctuationOrSymbol(before);
const afterIsPunct = isPunctuationOrSymbol(after);
const leftFlanking =
!afterIsSpace && (!afterIsPunct || beforeIsSpace || beforeIsPunct);
const rightFlanking =
!beforeIsSpace && (!beforeIsPunct || afterIsSpace || afterIsPunct);
const cjkPunctBefore = !!before && CJK_CLOSING_PUNCT_RE.test(before);
const wordAfter = !!after && WORD_CHAR_RE.test(after);
delimiters.push({
pos: cursor,
canOpen: leftFlanking,
canClose: rightFlanking || (cjkPunctBefore && wordAfter),
});
cursor += 2;
continue;
}
cursor += 1;
}
const stack: Array<{ pos: number }> = [];
const pairs: Array<{ open: number; close: number }> = [];
for (const delimiter of delimiters) {
if (delimiter.canClose) {
let openerIndex = -1;
for (let j = stack.length - 1; j >= 0; j -= 1) {
openerIndex = j;
break;
}
if (openerIndex !== -1) {
const opener = stack.splice(openerIndex, 1)[0];
pairs.push({ open: opener.pos, close: delimiter.pos });
}
}
if (delimiter.canOpen) {
stack.push({ pos: delimiter.pos });
}
}
if (pairs.length === 0) return normalized;
const insertPositions = new Set<number>();
for (const pair of pairs) {
const insideLast = normalized[pair.close - 1];
const afterClose = normalized[pair.close + 2];
if (!afterClose) continue;
if (
CJK_CLOSING_PUNCT_RE.test(insideLast) &&
WORD_CHAR_RE.test(afterClose)
) {
insertPositions.add(pair.close + 2);
}
}
if (insertPositions.size === 0) return normalized;
let result = "";
for (let idx = 0; idx < normalized.length; idx += 1) {
if (insertPositions.has(idx)) {
result += " ";
}
result += normalized[idx];
}
if (insertPositions.has(normalized.length)) {
result += " ";
}
return result;
}
export function fixCjkEmphasisSpacing(content: string): string {
const parts = content.split(/(^```[\s\S]*?^```|^~~~[\s\S]*?^~~~)/m);
return parts
.map((part, i) => {
if (i % 2 === 1) return part;
const blocks = part.split(/(\n\s*\n+)/);
return blocks
.map((block, index) => {
if (index % 2 === 1) return block;
return fixCjkEmphasisSpacingInBlock(block);
})
.join("");
})
.join("");
}